SURVEY OF THE BRITISH ISLES FOR THE EPOCH JANUARY 1. 1886. 
On tlie whole, then, we think the application of the correction for diurnal range is 
advantageous, but the correction for disturbance appears to be somewhat uncertain in 
its effect as applied to the field observations, and we have retained it chiefly for the 
sake of uniformity with our treatment of the other elements. 
The formulm used in the application of the corrections to the Horizontal Force and 
Hip were as follows : — 
If H', m' and H, m be the values of the Horizontal Force and the moment of the 
magnet deduced from the uncorrected and corrected observations respectively, and if 
(f) and ifj be the total increments of the Horizontal Force due to diurnal variation and 
disturbance at the times when the deflection and vibration experiments are made, 
H = H'-(c^ + ,^)/2; 
m = m {1 + — '/') / 2H']. 
As the disturbances of the Dips have to be deduced from those of the Horizontal 
and Vertical Forces, the calculations involved are rather troublesome. 
If dd be the change, expressed in minutes of arc, produced by increments dY and 
f/H in the Vertical and Horizontal Forces respectively, 
dd = adY - hdB., 
where a and h are quantities which depend on the Dip and Horizontal Force at the 
station. Tables were prepared in v/hich the values of these were entered for each 
complete degree between 6G° and 72”^, and for each tenth of a unit of Horizontal 
Force between 1’4 and 1’8. As they vary slowly, the value corresponding to any 
given Dip or Force could be readily determined, and the value of dO was thus 
deduced. 
Method of Tahidating the Observations. 
We have ventured to modify considerably the methods ordinarily adopted of pre¬ 
senting the results of a survey such as our own. It is often very difficult for any one 
who studies the records of magnetic observations to learn anything as to whether the 
instruments were in good or bad adjustment, or whether the actual observations were 
careful or careless. Sometimes superfluous data are given which supply no infor¬ 
mation on these points. Thus, the quantities log mX and log m/'K, or mX and m/X, 
which are sometimes tabulated in records of Force observations, add but little to their 
value. In an observatory, indeed, they ought to be nearly constant from month to 
month; but as, in both of them, the changes in the value of the moment of the 
magnet and those in the Force are mixed up with the error of experiment, it is not 
easy to draw any definite conclusion from them. To test the observations, we should 
compare quantities the difference between which ought theoretically to vanish. In 
the case of a survey, log rnX and log m/X must vary largely from station to station, 
and they are,, therefore, quite worthless as tests of the observations. 
